Considering the effects of vehicle downsizing and ridesharing service on urban transportation systems is very important due to the increasing population in urban areas and decreasing free space on urban roads and streets.These two issues have significant effects on various aspects of urban transportation services,e.g.,fuel consumption,emission,safety,economy,parking space,and sustainability.This thesis aims to investigate the influences of vehicle downsizing on traffic,so that we first categorize the passenger cars into three subsets,i.e.,small cars(hatchback cars without trunk),sports utility vehicles(SUVs),and standard cars(taxis and regular private cars),and measure their passenger car equivalent values by using the mean time headway method.Due to various weather conditions,we measure the passenger car equivalent(PCE)values on sunny days and moderate rainy days,respectively.The results show that the PCE values for small cars and SUVs are 0.87 and 1.26 on sunny days,and 0.87 and 1.31 on rainy days,respectively.Then based on the lower PCE of the small cars,it is essential to understand the effect of the vehicle size on the delay and level of service(LOS)at intersections.Therefore,we study the influence of the vehicle size on the performance of signalized intersections by using VISSIM simulation software.The simulation results of a case study in Hangzhou,China,show that(a)the reduction in the vehicle size can decrease the intersection queue length and delay if replacing standard cars and SUVs with small cars;(b)the average queue length can be reduced by more than 10 m with replacing all cars with the small cars.The regression result shows that with unit queue length decreased,the average delay can be reduced by 0.65 s.In our research,vehicle downsizing has a positive impact on LOS.As a case study,we investigate the impact of taxi downsizing on efficiency and sustainability in the taxi industry in Hangzhou.The use of hatchback cars as taxi vehicles makes the taxi size smaller,approximately 80 cm in length and 15 cm in width.With supplementary surveys with taxi passengers,a total of 1,110 taxi trips are observed at nine diversified locations.The results show that the trunk is not used in 88%taxi trips in Hangzhou,while 60%taxis only pick up one passenger,and 29%taxis pick up two passengers.Furthermore,a hatchback car with two-row seats is lighter and smaller than a regular taxi vehicle.It is estimated that each taxi downsizing results in annual decreases in the fuel consumption by 1,600 liters,CO by 311.9 kg,HC by 15.4 kg,and NOx by 8.9 kg,respectively.By analyzing GPS data of 7,081 taxis in Hangzhou,we present the network-wide taxi average speed is 23 km/h.Since the average speed of taxi is relatively low in urban areas,the use of small cars could provide the driver with the higher maneuver capability,while without increasing the safety risk of small cars.Besides,taxi downsizing influences economic efficiency;for instance,1,600 liters of fuel and $91 of maintenance costs can be saved annually.The results present that using small cars in replacement with regular taxis will increase the efficiency and sustainability of the taxi industry.Besides,ridesharing can also decrease the number of vehicles on the roads.Therefore,we present an ensemble learning approach for better understanding the ridesplitting behavior of passengers of ridesourcing companies who provide prearranged and on-demand transportation services.An ensemble learning model is a weighted combination of multiple classification models or weak classifiers to form a strong classification model.The goal of ensemble learning is to combine decisions or predictions of several base classifiers to improve prediction generalizability and robustness over a single classifier.This research employs the Boosting ensemble by growing individual decision trees sequentially and then assembling these trees to produce a powerful classification model.To improve the prediction accuracy of ridesplitting choices,we explore real-world individual-level data extracted from the on-demand ride service platform of DiDi in Hangzhou,China.Over one million trips of the four service types,i.e.,Taxi Hailing Service,Express,Private Car Service,and Hitch,are explored with descriptive statistics.A variety of features that may impact ridesplitting behavior are ranked and selected by using the ReliefF algorithm,such as trip travel time,trip costs,trip length,waiting time fee,travel time reliability of origins/destinations,and so on.The ensemble learning approach was presented to train both the full(with all features)and reduced(with only important features)ridesplitting models by appropriately fusing the estimations obtained by multiple weaker but efficient classification decision trees.The results show that the Boosting ensemble trees with full features returned classification errors of 0.177 on the training set,and 0.225 on the test set,respectively,while the Boosting ensemble trees with selected features returned comparable results without losing too much accuracy.In summary,for reducing the allocated road space to cars,this thesis presents the benefits of small car usage from the perspectives of energy-saving,pollution diminishing,and effects on traffic.If governments and urban planners encourage people to use ridesharing services and small cars,urban transportation systems will step toward sustainability. |